Distribution Theory and Simulations for Tests of Outliers in Regression

 

 Anthony C. Atkinson Marco Riani
The London School of Economics, Dipartimento di Economia,
London WC2A 2AE, UK UniversitÓ di Parma,
 UK Italy
 a.c.atkinson@lse.ac.uk mriani@unipr.it

Abstract


Three straightforward methods using truncated simple samples are described for approximating the distribution of the test statistic for multiple outliers in regression. The simulations also proved a powerful method of calibrating pointwise inference for simultaneous tests for an unknown number of outliers. Analysis of data on fidelity cards reveals an unexpected group of outliers.
 

 

To download the .pdf version of the paper click here (link to the Dept. of Stat. of the London School of Economics)

 

To download fidelity cards data click here (ascii format)

 

Description of the variables

x1 (first column of the file ) = number of visits to the supermarket in the six months period

x2 (second column of the file) = age of the customer

x3 = (third column of the file) = number of members of the customer's family

y = (fourth column of the file) = amont in euros, spent at the shop over six months


Last modified 04/06/2009 21.08.41